Infrastructures Development, Environmental Quality and Economic Growth in Nigeria

The Earth as a planet supports human life, living and activities, attracting extensive and intensive socioeconomic influences on the economy. Such activities like infrastructures development exerts increasing and divers environmental quality concerns and hence the economic growth. While these variables appear interrelated due to many factors including population growth, urbanisation, etc. However, the relationship between infrastructures, environment and economic growth is not largely known especially in Nigeria. This study therefore investigated their relationship using time series data between 1990 and 2019 by adopting Co-integration estimation technique through the Bound test approach of auto regressive distributive lag method using percentage share of building and construction sector of gross domestic product (GDP), carbon dioxide, population growth GDP growth rate, etc. as variables. The study revealed that the infrastructures development and environmental quality explain economic growth and have both short and long run relationships while specifically population growth and agriculture, forestry, fishing, value added variables are positively significant to economic growth. The findings evidences of both short and long run relationships among the variables are significant and it is consequently recommended that new roles for infrastructure sets and production processes should consider environmental quality mindsets to achieve positive green economy outcomes in Nigeria. JEL Classifications: O18, O44, Q5


Introduction
It is an established scientific fact that the Earth is the only planet of the solar system that substantially supports human life and living. The dynamics of human living on earth have also attracted extensive and intensive socioeconomic activities while interacting with the environment-ecosystem. These activities

Statement of the Problems
Ideally, infrastructures development underscores economic growth and defines the economy of economies but not without environmental sustainability concern. For example, road infrastructures facilitate efficient movement of goods and services but their construction affects biodiversity; water supply dam development improves agriculture, health and sanitation but makes the environment vulnerable to flooding, pollution, deforestation, etc. The infrastructures-environment nexus has been adduced to many factors not limited to industrialisation, increasing population growth, urbanisation, etc. (FGN 2017;Sulaiman and Abdul-Rahim 2017;. Nevertheless, the nexus creates positive and negative externalities for economic growth in developed and developing economies (Maliszewska and Mensbrugghe 2019) not limited to improved income per capita and output, yet decreasing agricultural output, increasing air pollution, health challenges, etc. There is therefore tension between the need for infrastructure development for economic growth and the goal of ensuring environment quality. From Nigerian experience, studies on the infrastructures, environment and growth nexus seldom enjoy mention, though there are few related efforts like GHG emissions and economic growth (Issaoui et al. 2015), climate change, water availability and agriculture (Metu et al. 2016), environmental impact assessment (EIA) for capital spending on infrastructures development for growth (Babatunde 2018), etc.

Research Questions
This noticed of infrastructures-environment and growth nexus in literature informed this study especially with respect to providing answers to these questions: Does environmental quality and or infrastructures development explains economic growth in Nigeria? What is the nature of relationship between infrastructures development, environmental quality and economic growth in Nigeria?

Objectives of the Study
This study therefore intends to examine the relationship between infrastructures, environment and economic growth in Nigeria.

Justification for the Study
The essence of this study is on the contribution of infrastructures and environment to growth of economies, specifically that infrastructures construction exerts positive and negative externalities on the environment and in turn influences on other economic sectors in developed and developing economies like Nigeria.

Structure of the Study
The rest of this study is organised as follows: literature review, research gap, methodology and the last section provides conclusion and policy recommendations.

Conceptual Review
Infrastructures, in the views of Buhr (2009), Fasoranti (2016, Regan (2017) and Losos et al. (2018) are a set of heterogeneous, essential, hard and soft, social and economic assets. They are characteristically intangible, physical but capital intensive public goods of long life cycle (Flyvbjerg 2014;Iyortyer 2017) and factor inputs of overall modern economy, long-term capital appreciation, welfare especially in the developing economies. van der Putten (2016), Chakrabarti (2018) and World Bank (2019) argued that though infrastructures define economy of economies but recognised possible hindrances and constraint to ambitious infrastructure investment and growth as range of associated environmental and financial concerns, geopolitics, public policy, socio-cultural spatial linkages, etc.
The environment substantially supports human life and living activities that promote growth but is vulnerable to intensive and extensive risks, for example, pollution (air, water, soil and radioactive) (Ohiare 2015). However, some alternative routes and attempts are being developed to reduce these environmental risks and their economic consequences such as, exploring renewable energy sources, encouraging green housing construction methods and materials, developing electric-powered transport system, improving road networks and traffic, etc., according to Olanipekun (2016), Losos et al. (2018), Meng and Han (2018) and Li et al. (2019).
Though economic growth has been severally construed (Adelakun 2011;Edeme 2018;Fasoranti 2016) but central to the views that economic growth is the sustained and greater increase in output (or income) as a derivative of efficiency and amount of inputs factors in an economy over a long time period usually expressed in GDP. The changes in nominal GDP from one year to the next across all economic sectors expressed in percentage reflect economic growth rate and serve as common bases for measuring performance and strength of nation's economy (Fasakin and Jegede 2018). For example, the Nigerian economy grew at an average of 2.45% GDP between years 1999 and 2017), specifically with 5.52% GDP in year 2000, 6.8% GDP in 2005 through 9.54% in 2010 but slowed to 2.79% in 2013 and contracted by −1.5% in 2016 (NBS 2016) but recovered with average of 1.95% since 2107/2018 fiscal years (NBS 2017). Economic growth accounting includes intensity and scale of infrastructures spending but largely exclude production from the informal market and environmental vulnerabilities, etc. in Nigeria (Babatunde 2018;Teo et al. 2019). Hence, growing economy at expense of environmental quality and/or huge infrastructural development impairment is abnormal.
According to Rapu et al. (2015), FGN (2017), Sulaiman and Abdul-Rahim (2017) and Wang et al. (2018), quest for better standards of living and increased population drive increase in human activities such as infrastructure development, natural resources exploration, industrialisation, GHG emissions, pollution, climate change and food insecurity, etc., with long-run impact on the economy. This is evident in the Nigeria's population growth rate at about 3%, estimated 17 million housing deficit, 66.9% land use change effects for urbanisation, huge disruption to the ecosystem, social and economic costs that influence market failures (Maliszewska and Mensbrugghe 2019). Figure 1 compares economic growth proxy by GDP growth rate and carbon dioxide (CO 2 ) emission in Nigeria, and underscores the common argument that environment impairment increases with increase in industrialisation leading to economic growth, however, reversible with public regulatory policies.
Globally, regulatory policies have favoured the concept of green economy across sectors (World Bank 2019). These policies emphasise sustainability concepts such as growing the economy yet improving environmental quality by exploring the routes of fight against climate change, environmental taxes, carbon pricing, renewable clean energy sources, innovative energy incentives, etc. (Zeleňáková et al. 2018). In Nigeria, green economy policies such as EIA Act (1992) and Sustainable Development Goal (2015) indulge infrastructures planning stage nexus with the environmental quality and protection on certain class of economic infrastructures projects for which EIA report must be conducted and approved before implementation. Additionally, reducing negative effects of infrastructures development through effective preventive maintenance, motivation and innovation via alternative mode of transportation, etc.

Theoretical Review
Earlier thoughts on input-output assumption related to economic growth (Y') have been severally considered and agreed as a production function by the classical (Smith 1776) centred on rate of population growth or labour (l) and particularly land (d) and its resources use (agriculture, minerals, etc.) as disruptions that influence growth simply expressed as: The neoclassical growth (Solow 1956) theorists further argued that the use of land (d), labour (l) inclusive of capital (k) (e.g., financial resources) with influence of technological progress (A) will result in a longrun output growth in an economy, simply expressed thus: Here A shows productivity level of technical progress (like infrastructures development and machines) that facilitates efficiency of other factor inputs for improved output and hence growth (Babatunde 2018) assuming constant return to scale, except on capital.
From the foregoing models, it shows that environmental factors are taken for granted and seldom accounted for in growth modelling while knowing that natural resources input factors are in high demand but scares in supply processes since they exert environmental risks as negative externalities in the production process (Maliszewska and Mensbrugghe 2019;Wolde 2015). However, Environmental Kuznets Curve is widely believed to explain the economic growth-environmental risk relationship emphasising that as per capita income rise, pollution and other forms of environmental degradation would rise 1st and then fall in an inverted-U pattern. Further, critiques have emphasised that environmental risks nexus economic growth negatively (Cumming and Cramon-Taubadelb 2018) especially in developed economies in modern growth theories. Therefore, environment impairment (e) assumes to increase with fixed capital (K) (infrastructures development) and in turn may increase/decrease economic growth: Therefore modern growth modelling assumptions approach to input-output assumption related to economic growth (Y'), is appreciably inclusive thus: Shuaibu and Oyinlola (2013) established no causal link between CO 2 emissions and energy consumption to economic growth due to structural shifts. Using time-series analysis on annual data from 1981 to 2013, and adopting residual-based cointegration test with a structural break, it was found that there is current account sustainability in Nigeria and structural changes were not very potent during the period under consideration. This implies that the Nigerian economy complied with the Inter-Temporal Budget Constraint (IBC) hypothesis, suggesting that exports could actually finance imports. Sahrir et al. (2014) studied 'Environmental and health impacts of airport infrastructure upgrading: Kuala Lumpur International Airport' and found that significant environmental concerns are noise and air quality. Using field survey, sampling and analysing noise level and airborne particles of the outdoor and indoor at the airport construction and adjacent sites thus, recommended increased construction site environmental sensitivity methodology and land use. Chingoiro and Mbulawa (2016) examined economic growth and infrastructure expenditure in Kenya. Using annual data on GDP growth rate, infrastructure spending and labour for the period 1980-2013 and employed Granger causality approach, it was found that there is bidirectional flow of causality between economic growth and infrastructure, recommending that the government should commit more funds towards developing infrastructure in the short term. Alege et al. (2016) investigated pollutant emissions, energy consumption and economic growth in Nigeria to find the direction of causal relationships among emissions, energy consumption and economic growth using annual time series data for the period 1970-2013 and adopting Johansen maximum likelihood cointegration and Granger causality tests. While the result showed existence of cointegrating vector between the variables at long run that is, fossil fuel enhances carbon emissions in the atmospheric, it also shows existence of unidirectional causation from fossil fuel to CO 2 emissions and GDP per capita.

Empirical Reviews
Cumming and Cramon-Taubadelb (2018) studied linking economic growth pathways and environmental sustainability by understanding development as alternate social-ecological regime by analysing the red loop-green loop model that proposes growth defined by Human Development Index is explained by social-ecological interactions shift proxied by populations growth, traditional cultural practices and natural resources per capita GDP using Pearson's correlation and found that environmental and ecological sustainability are increasingly significant in economic growth but with little effect in development practices. In Nigeria, Mba (2018) adopted primary and secondary sources in a study on 'Assessment of environmental impact of deforestation in Enugu, Nigeria', and concluded that major contribution to deforestation in this area is urbanisation and industrial development caused by population increase and recommended that government ensure public awareness, monitoring and enabling laws to deter the trend. Babatunde (2018) investigated government spending on infrastructure and growth using primary and secondary data from 1980 to 2016 in Nigeria and adopted weighted least square model and found that government spending on transport and communication, education and health infrastructure has significant effects on economic growth in Nigeria. Li et al. (2019) examined the EIA of transportation infrastructure life cycle using the three phases of construction, maintenance and repair and demolition of the fast track transportation project in China, by taking measurement of materials used and the energy consumed as environmental emissions, and found that construction, demolition and maintenance phases have environmental impact in descending order especially in the use of steel material. Sturup and Low (2019) examined 'Sustainable development and mega infrastructure: An overview of the issues' in China and found a strong relationship between global ecosystem and mega infrastructure development, asserting a balanced sustainability concepts for physical development of infrastructures. Odugbesan and Rjoub (2020) assessed the 'Relationship among economic growth, energy consumption, CO 2 emission and urbanisation: Evidence from MINT (Mexico, Indonesia, Nigeria, and Turkey) countries' employing the auto regressive distributive lag (ARDL) Bounds test approach, and the result revealed that the energy-growth hypothesis has unidirectional causality from energy consumption in Nigeria and Indonesia amongst others and show a long-run relationship with economic growth, energy consumption and CO 2 emissions. Urhie et al. (2020) examined economic growth and environmental impacts relationship by adopting moderated mediation model to assess the cyclical effects of these economic relationships of air pollution and health outcomes and found significant interaction between air pollution and government expenditure on health performance hence asserting that environment friendly production and consumption pattern minimise environmental hazards and recommended that preventive public policies to adverse health outcomes on manufacturing firms.

Research Gap
Some of the studies reviewed establish that infrastructures and/or environmental concerns explain growth using different estimation techniques. However, none of the studies reviewed established infrastructures development-environment quality nexus influence on economic growth particularly as related to Nigeria indicating paucity of studies in this area. Besides, very few of the studies reviewed like Odugbesan and Rjoub (2020) adopted ARDL estimation technique. This study intends to fill this gap and contribute to the knowledge area qualitatively and quantitatively using the variables from Nigeria.

Theoretical Framework and Data Sources
In order to achieve the objectives of this study, the quantitative analysis adopted a production function model with the econometric model specification from the work of Sulaiman and Abdul-Rahim (2017) with modifications. The modification involves the derivative decomposition of the exogenous parametric variables as related to economic growth. This study framework is premised on emerging modelling of economic growth as a function of climate change (environmental sustainability) induced by physical development activities. This may correspond GDP growth as an economic measure and increase in environmental risk like pollution, deforestation GHG emission via deliberate economic activities of infrastructures development, to achieve and influence desired economic and social macro-economic objectives across economic sectors (Issaoui et al. 2015). The linear function specified for the framework is: Where, e = (K, L) as in traditional production function defined as capital and labour respectively, and h depicts the rate of infrastructures development activities. For this study, economic growth as the endogenous variable is proxy by GDP growth rate (GDPr), while population growth rate (PPMr), BCS share of GDP, CO 2 from transport, electricity and heat production (% of total fuel combustion) (CTE) and agriculture, forestry and fishing, value added (annual % growth) (AFF) are the exogenous variables. A standard log-linear transformation functional relationship was employ between endogenous and exogenous variables in order to stabilise the variance of the variables or avoid heteroscedasticity. Time series secondary data (AFF) and (CTE) are obtained from World Bank World Development Indicators while (GDPr), (BCS) and (PPMr) are obtained from Central Bank of Nigeria (CBN) Statistical Bulletins and CBN Economic Reports spanning years 1990-2019 and are used for the study.

Model Specifications
The endogenous model often specified for testing or explaining the effects of the independent variables on dependent variable is expressed in estimation equation or function. The linear function specified for the estimation in this study is: Where, GDP growth rate (GDPr), Environment = f(CTE), Infrastructure = f(BCS), Land = f(AFF) and Labour = f(PPMr).
Then, the overall function is mathematically expressed as follows: The set of variables in this study were used because of their conjectural strong influence on economic growth, as qualitatively discussed, with respect to infrastructure development-environment quality nexus. The Equation (7) is further transformed into an econometric model as follows: Where, GDP growth rate (GDPr), population growth rate (PPMr), building and construction sector (BCS) share of GDP,CO 2 from transport, electricity and heat production (% of total fuel combustion) (CTE) and agriculture, forestry, fishing, value added (annual % growth) (AFF); while β 0 = intercept term, β1 = coefficient of CTE, β 2 = coefficient of AFF, β 3 = coefficient of BCS, β 4 = coefficient of PPMr and u t = stochastic or disturbance term. On a priori ground the various theoretical expectations explained above are: Furthermore, since the model variables are not in the same unit scale, this study specifies its model in a simple log-linear form by taking the partial natural logarithm of some variables in the Equation (9). This gives the below equation: Where, β 0 = lnA t− 1 or the intercept with lag and u t remain the disturbance error term. Equation (7) is the long-run relationship infrastructures development-environment quality and economic growth nexus.

Estimation Technique
Adebiyi (2003) opined that macro time series data requires unit roots and co-integration tests before a structural relationship is estimated. On this basis, estimation process for this study involved unit root test of the series data adopting only the Augmented Dickey-Fuller (ADF) technique (at level and difference) to estimate the series stationarity, which was found to be in mixed orders of integration. This is followed by co-integration analysis employing Bounds test approach of the ARDL method in order to ascertain the nature of relationships of the series variables in the model in Equation (9), which was found to be of long run nature, as first developed in the work of Pesaran et al. (2001). The ARDL method is a type of ordinary least square model, useful and applicable to small samples size with both non-stationarity and stationarity time series at varied integration levels. It further estimated the short run relationship equilibrium by exploring unrestricted error correction model (ECM) thus: Equation (10) is a linear model of appropriate lag level length of one of the variables used to estimate the both short run effects, particularly with the 2nd part with the summation signs (Σ) and (∆) representing the estimation of the short run dynamic based on Akaike information criterion, resulting in ECM. The short-run effects are revealed by the estimates of coefficients attached to first differenced variables. Following, the post estimation diagnostic statistics tests of the infrastructures developmentenvironment quality and economic growth nexus such as descriptive statistics, multicollinearity, serial correlation, heteroscedasticity and cumulative sum of squares of recursive residuals for model variables were carried out to ascertain the reliability of the parameter estimates.

Results and Discussions
Only ADF technique was employed for the unit root test on the series data to examine their stationarity or otherwise at different orders of integration and critical value at 5% level of significant adopting Equation (9) in order to avoid the properties of stochastic error terms and spurious results in the model. The result is shown in Table 1.
The condition for stationarity is that the absolute value of the ADF statistic must be greater than that of critical value of corresponding variable and p-value is less than 5%. From Table 2, the result of the series variables at level shows that both logCTE and PPGr are not stationary indicating existence of unit roots. However, at first difference, all the variables were stationary. In all, there is a mix of orders of integration at I(0) and I(1) amongst the series variables underscoring the necessary condition for estimation using ARDL model technique. Table 2 presents the result of ARDL Bounds test for co-integration.

Co-integration Test
From Table 2, the F-statistic is compared with the critical value at only 5% significance level, and shows that the calculated F-statistic value is 6.821, far above the upper (4.01) or (5.06) and lower (2.86) or (3.74) bounds of the critical values, indicating that cointegration exists. Therefore, the null hypothesis of no cointegration is rejected at 1% and 5% significance levels and hence there is a long-run relationship among the variables in the model. This finding infers that economic growth is very responsive to changes in these exogenous variables specifically infrastructures development-environment quality nexus and sustainable in the long run.
With the established result of co-integration above, both long and short run (dynamic) models were estimated using ARDL co-integrating and long run form regression technique in determining the coefficient of the ECM as shown in Table 3.
The result of the cointegrating regression in Table 3 showed that the overall estimated values of the model are good with R 2 = 0.7600 or 76% variation in GDPr (dependent variable) is explained by PPGr, LNCTE, BCSr and AFF (independent variables) while the remaining 34% would be explain by other variables not included in the model. The Durbin-Watson value (1.97) implies no autocorrelation in the model, while the Pro(Stat) value (0.000007) indicates that the overall model is statistically significant at 5% level of significance. Furthermore, all the independent variables in the long run estimate (PPGr, LNCTE, BCSr and AFF) have positive relationship with GDPr indicated by their coefficients which are greater than 0 and hence consistent with the a-prior expectation of the study. This result underscores the study by Odugbesan and Rjoub (2020). In the short run however, the result showed that coefficient of D(PPGr) (12.795) and D(AFF) (0.1809) with prob stat (0.0245) and (0.0007) respectively are significant at 5% and positively correlated. There positive signs agree with Sulaiman and Abdul-Rahim (2017) and Wang et al. (2018), that increased population growth drives human activities such as infrastructure development, agriculture, etc., leading to output and better standards of living. The CointEq(−1) value (−0.892) which measures the speed of adjustment flow is both negative and significant indicating that 89% of the errors are corrected and in approximately 26 years for the economy to attain equilibrium. This supports the emphasised long environmental risks effects on economic growth especially in developed economies and as in the postulation modern growth theories (Cumming and Cramon-Taubadelb 2018).

Post Estimation Diagnostic Tests
The post estimation diagnostic tests such as descriptive statistics, multicollinearity, serial correlation, heteroscedasticity and cumulative sum of squares of recursive residuals for model variables were carried out thus.  The descriptive statistics of the variables presented in Table 4 A show the mean, median and standard deviation statistical values of the variables.
The results in Table 1 show the mean growth values of the variables and indicates that AFF (6.094) has the highest growth rate while PPGr has the least values of growth. The highest degree of standard deviations is associated with AFF (9.496) while PPGr (0.079) has the lowest amongst others variables. These indicate that the probability of obtaining positive outcome is high. Table 4 B shows a test of multicollinearity amongst the variables in the model in whether the explanatory variables are highly interrelated. Where any value of the variables has centred variance inflation factors greater or equal to 10, then there is multicollinearity, where it is less, then there is no multicollinearity. From Table 4 A, it show that none has value greater than 10, indicating that there is no multicollinearity variables.

Serial Correlation Test
In Table 4C, Breusch-Godfrey Serial Correlation Lagrange multiplier (LM) test model was used for serial correlation test. The Prob. Chi-Square value of (0.3089) obtained, which is not significant at 5%, shows that there is problem of serial correlation in the model and hence cannot reject the null-hypothesis of no auto correlation.

Heteroscedasticity
Using the Breausch-Geofred LM test, as shown in Table 5, the probability chi-squared value of 0.0105 (1%), which is significant at 5% shows there is no problem of heteroscedasticity or ARCH effect in the model, hence rejecting the null-hypothesis. Figure 2 is the CUSUM test and the CUSUM of Squares plots showing that the cumulative sum blue lines lies within/between the 5% critical lines (straight bounded upper and lower red), suggesting that the coefficients of the parameters in the ECM at short-run and the long-run estimates are not only stable over the period of the study but also do not suffer from any structural instability.

Conclusion
This study employed an ARDL approach to cointegration in a recursive form to assess the nature of the relationship between infrastructures development-environmental quality nexus and economic growth in Nigeria using variables including GDP growth rate, population growth rate, BCS, CO 2 from transport,   electricity and heat production and agriculture, forestry and fishing, value added as variables. The behaviour of the variables explained their significance in the long-run and short-run on the economic growth while the diagnostic tests conducted for the models revealed the good fit and satisfactory classical linear regression requirements. However, the diversity and intensity of infrastructures development-environmental quality nexus across economies will require new roles for infrastructure sets and production processes that will consider environmental quality mindsets to achieve positive green economy outcomes in Nigeria. In addition, government should pursue policies and programmes in infrastructures like renewable energy projects that will be eco-friendly yet enhancing economic activities and growth in Nigeria.